# The Unbearable Weight of Generating Artificial Errors for Grammatical Error Correction

## TL;DR

This paper explores the use of neural models to generate artificial grammatical errors for training error correction systems, analyzing how different models and data sizes impact correction performance.

## Contribution

It introduces a neural error generation approach and systematically compares it with rule-based methods, highlighting its effectiveness for grammatical error correction.

## Key findings

- Neural error generation improves correction accuracy.
- Model choice significantly affects error realism.
- Data size influences correction performance.

## Abstract

In recent years, sequence-to-sequence models have been very effective for end-to-end grammatical error correction (GEC). As creating human-annotated parallel corpus for GEC is expensive and time-consuming, there has been work on artificial corpus generation with the aim of creating sentences that contain realistic grammatical errors from grammatically correct sentences. In this paper, we investigate the impact of using recent neural models for generating errors to help neural models to correct errors. We conduct a battery of experiments on the effect of data size, models, and comparison with a rule-based approach.